794 TRUEBLOOD. 



that the effect of moisture is negligible — not, however, that moisture 

 is necessarily absent. 



The presence of varying amounts of moisture in supposedly dry 

 steam has been the occasion of so considerable an error in the results 

 of at least one experimenter that it was deemed advisable, in the 

 present work, to provide with some care against the entrance of error 

 from this source. The precaution^ taken have already been described. 

 They consisted in attention to the design of the primary superheater, 

 in operating the secondary superheater as a cooler, and in providing 

 strainers. The object of strainers is in part to get rid of moisture by 

 the operation of throttling, on the principle embodied in the ordinary 

 throttling calorimeter used for determining the quality of moist steam, 

 and in part to break up particles of moisture into small drops, from 

 which further evaporation is less likely to occur. The only actual 

 tests on the efficacy of these precautions are given in the results ob- 

 tained with plugs Ul and U2. The arrangements of the apparatus 

 in these two cases differed only in that a copper gauze strainer was 

 used with Ul and an alundum strainer with U2. A glance at the 

 curves of Figs. 12 and 13 will show that this test indicates no appre- 

 ciable wet steam effect. In fact, such discrepancy as appears is in 

 the wrong direction for explanation on the supposition that wet 

 steam was the cause of it, since the alundum strainer was far more 

 effective than the gauze strainer. 



The low values of fx obtained by extrapolation with the two I^-type 

 plugs are mainly due to heat-leakage into the annular space surround- 

 ing the radial flow plug. The steam is here spread out into a rather 

 thin layer not effectively insulated from the bath, and as its tempera- 

 ture has been depressed, before it reaches this space, by throttling 

 in the mixing chambers at the two ends of the cross channel, the 

 opportunity for heat-leakage is excellent. The need of locating the 

 high side thermometer in close proximity to the radial flow plug is 

 well illustrated by these results. Although the set-up is similar to 

 that used with the axial flow plug A4, the results do not form a fair 

 basis of comparison betw^een the two types of plug, since the chief 

 advantage of the radial flow plug is sacrificed bv the arrangement. 



1 ' ; 



In both of the U-type plots of yu' vs. , there is evidence of curvature, 



which, if actual, would lead to larger ijl's than are given by the straight 

 line extrapolation. Whether this curvature is real or not has no 

 particular bearing on the evidence of the plots as to the question of 

 wet steam, and the plots have practically no other value than this. 



